Clozapine, an atypical neuroleptic, is more effective for treatment- resistant schizophrenia than typical neuroleptics previously available and is less likely to cause extrapyramidal side effects. Its unusual therapeutic profile is matched by a broad-spectrum pharmacologic profile. This combination of clinical and pharmacological effects makes it a useful probe for studying biochemical factors involved in the response to anti- psychotic drugs. However, despite its potential clinical and research value, there are two problems with its use: it is highly toxic and very expensive. These problems suggest that a predictor of response could be of value. Preliminary findings from our group and others have revealed that pretreatment (neuroleptic-free) plasma and cerebrospinal fluid (CSF) measures of biogenic amines and their metabolites may be of value in predicting response to clozapine; preliminary findings from our group and others also suggest that, during treatment, changes in these measures may be correlates of clinical response to atypical and typical neuroleptics. Utilizing a 6 week drug-washout period, we propose to do a l2 week double- blind study of the atypical neuroleptic clozapine (CLOZ) and the typical neuroleptic chlorpromazine (CPZ) in 90 treatment-refractory schizophrenic patients (to obtain a final sample of 48). The study will utilize plasma and CSF measures of biogenic amines and their metabolites to achieve the following specific aims: (l) To confirm and extend preliminary findings suggesting that pretreatment plasma levels of homovanillic acid (HVA) and possibly norepinephrine (NE), as well as the ratio of CSF HVA to CSF 5- hydroxy-indoleacetic acid (5HIAA), predict clinical response to CLOZ; (2a and 2b) to confirm and extend our preliminary findings suggesting that decreases in plasma levels of HVA and 3-methoxy-4-hydroxy-phenylglycol (MHPG) during treatment with CLOZ correlate with clinical response to CLOZ; (2c and 2d) to explore the relationships between patterns of changes in plasma and CSF measures of biogenic amines and their metabolites and clinical response to CLOZ treatment; (3a) to confirm and extend our preliminary findings on the effects of CWZ treatment on plasma levels of HVA, MHPG, dopamine (DA) and NE; and to explore the comparative effect of CPZ treatment on these substances; (3b) to explore the Comparative effects of CWZ and CPZ on CSF measures of DA, NE, HVA, MHPG, 5HIAA and HVA/5HIAA; and (4) to explore the relationships between clinical symptomatology and plasma and CSF measures of biogenic amines and metabolites at the end of the drug-washout period.